A Technical Consultation on
the Safety of Hepatitis B
Vaccines
Viral Hepatitis Prevention Board
Geneva, 28-30 September 1998
Report of a meeting organized
by the Viral Hepatitis Prevention
Board
by ANDY HALL and PIERRE VAN DAMME
on behalf of the Viral Hepatitis Prevention
Board
World Health Organization
WHO Collaborating Centre for Prevention
and Control of Viral Hepatitis
University of Antwerp
Viral Hepatitis Prevention Board
Centre for the Evaluation of Vaccination
WHO Collaborating Centre for Prevention and Control
of Viral Hepatitis
Epidemiology and Community Medicine
University of Antwerp
Universiteitsplein 1
B-2610 Antwerp
Belgium
TABLE OF CONTENTS
1.Introduction ................................................................................................................................. 3
2. Hepatitis B ................................................................................................................................. 3
2.1. Scope of the issue ........................................................................................ 3
2.2. Hepatitis B vaccine ....................................................................................... 4
2.2.1. Overview of assessment of vaccine safety .................................. 4
2.2.2. Vaccine safety and causality ........................................................ 5
3. Auto-immune diseases, including MS ....................................................................................... 6
3.1. Immunological mechanisms and molecular mimicry.................................... 7
3.2. Animal models and demyelination ................................................................ 8
4. Available data on safety of HB vaccines.................................................................................... 8
4.1. Literature and clinical trials ........................................................................... 8
4.2. Data from the USA ....................................................................................... 9
4.3. Data from Canada ...................................................................................... 10
4.4. Data from Italy ............................................................................................ 10
4.5. Data from Industry ...................................................................................... 10
5. Review of epidemiological studies in progress ........................................................................ 10
5.1. France ........................................................................................................ 10
5.2. UK............................................................................................................... 11
5.3. USA ............................................................................................................ 11
5.4. Industry ....................................................................................................... 11
5.5. Summary ................................................................................................... 12
6. Conclusions and consensus .................................................................................................... 13
Annexes:
Annex 1: list of participants .......................................................................................................... 15
2
1.
Introduction
In recent years, several case reports have raised concerns that hepatitis B (HB) immunization may
be linked to new cases or reactivation of multiple sclerosis (MS) or other demyelinating diseases
within a period of 2 to 3 months after vaccination . Articles published in the media stating a link
between HB immunization and MS have further fuelled the worry over the safety of the vaccine.
Although there is no epidemiological evidence showing that MS or flare-ups of MS occur more
frequently in recipients of HB vaccine, these concerns persist and have led to a reduction in the
uptake of HB vaccine in some countries, particularly in France.
Since many countries use HB vaccine in their immunization programmes, a thorough scientific
evaluation was warranted. The Viral Hepatitis Prevention Board (VHPB), whose activities are
incorporated into the World Health Organization (WHO) Collaborating Centre on Prevention and
Control of Viral Hepatitis, called a technical consultation on the safety of HB vaccines to review
accumulated data and policy implications. This consultation took place in Geneva from September
28 through September 30, 1998 and brought together representatives from national public health
and regulatory authorities, academia, the hospital sector, the pharmaceutical industry and the World
Health Organization. Participants included experts in the fields of public health, epidemiology,
immunology, neurology and pharmacology (see annex 1).
The primary objective of this meeting was to update the WHO statement regarding hepatitis B
vaccine and multiple sclerosis that had appeared in the Weekly Epidemiological Record in 19972.
The specific objectives of the meeting, on which this report is based, were to:
 summarise available data on the safety of HB vaccines;
 review systems for registering adverse events;
 review epidemiological studies conducted to determine the hypothetical relationship between HB
vaccines and neurological disorders;
 achieve a consensus on the safety of hepatitis B vaccine programmes.
2.
Hepatitis B
2.1. Scope of the issue
Hepatitis B (HB) is one of the major diseases of mankind and is preventable with safe and effective
vaccines. More than 2,000 million persons worldwide have serologic evidence of past or current HB
virus infection and there are more than 350 million chronic carriers of the virus at high risk of death
from cirrhosis and liver cancer, diseases which kill almost 1 million persons a year .
Although it is imperative to determine if there are any serious side effects related to hepatitis B
vaccine, it is also important to put the concerns over the safety of the HB vaccine in context,
weighing the real risk of HB infection and disease in non-immunized individuals against the
hypothetical risk of MS following HB immunization.
In the USA, for instance, the Centers for Diseases Control and Prevention (CDC) estimates (on the
basis of death certification data and specific studies) that of the 32,000 deaths from chronic liver
disease per annum (in the USA), 54% are due to hepatitis viruses B and C infections; 4,480 of these
chronic liver disease deaths are due to hepatitis B virus. Hepatocellular carcinoma deaths due to
Gout O, Lyon-Caen O. Sclérose en plaques et vaccination contre le virus de l'hépatite B. Rev Neurol 1998; 154: 205-7.
Anonymous. Lack of evidence that hepatitis B vaccine causes multiple sclerosis. Wkly Epidemiol Rec 1997; 72: 149-52.
Van Damme P, Kane M, Meheus A. Integration of hepatitis B into national immunization programmes. BMJ 1997; 314:
1033-36.
3
hepatitis B add another 3,000 deaths per annum. The annual costs of this infection in the USA are
estimated at US$ 658 million.
2.2. Hepatitis B vaccine
HB vaccines have been available since 1982 and more than 1 thousand million doses have been
used. Approximately 100 countries, consistent with World Health Organization policy, have added
HB vaccination to their routine childhood immunization programmes and most countries have in
addition adopted a policy to vaccinate adults at increased risk.
HB vaccine is considered one of the safest and least reactogenic vaccines. The vaccine is
approximately 95% effective in preventing the HBV chronic carrier state, and direct reduction of liver
cancer has already been documented in immunized children. The use of hepatitis B vaccine, both
plasma-derived and recombinant, has resulted in dramatic reductions in the prevalence of the
carrier state in many areas4 5. In areas without significant perinatal transmission the carrier rate has
been reduced to less than 1% from levels of 15-20%. In areas where perinatal transmission is
important (e.g. China) the reduction has been to less than 2% of the vaccinated population 5. There
is now clear data from Taiwan showing that hepatitis B vaccination reduces the incidence of liver
cancer in children6 7. It is evident that the benefit of HB vaccination outweighs any claimed risk in all
age groups.
2.2.1. Overview of assessment of vaccine safety
Several methods for assessing vaccine safety are in use. These include:
 Post-marketing surveillance – either passive or active;
 Reported case reports and case series;
 Data linkage methods.
In the USA, the Food and Drug Agency Center for Biologics Evaluation and Research (CBER) is
responsible for licensing and monitoring biological products such as vaccines, biological therapeutic
agents, blood and blood products. It carries out monitoring by collaboration with manufacturers in
the design and analysis of post-licensure studies and with the CDC in the Vaccine Adverse Events
Reporting System (VAERS) (passive surveillance) .
Although post-marketing surveillance, case reports and data linkage are useful in determining
vaccine safety and identifying adverse events associated with vaccination, certain drawbacks are
inherent in each:
Kane M. Status of hepatitis B immunisation programmes in 1998. Vaccine 1998; 16 (suppl.): 104-8.
Kane M. Hepatitis B control through immunisation. In: Rizzetto M, Purcell R, Gerin J, and Verme G, eds. Viral Hepatitis
and Liver Disease. Turin: Edizioni Minerva Medica, 1997: 638-42.
Chang MH, Chen CJ, Lai MS, Hsu HM et al. Universal hepatitis B vaccination in Taiwan and the incidence of
hepatocellular carcinoma in children. N Engl J Med 1997; 336: 1855-9.
Hsu HM, Lu CP, Lee SC et al. Seroepidemiologic survey for hepatitis B virus infection in taiwan: the effect of hepatitis B
mass immunization. J Infect Dis 1999; 179: 367-70.
It should be noted that in the USA any adverse event that has been reported and is possibly linked to a vaccine is listed in
the package insert because manufacturers believe this is an important element of openness in their relationship with
customers, whilst accepting that causality has not been established for many of the listed events.
4

Vaccine safety surveillance differs from drug surveillance in that the recipients of the vaccine are
usually healthy and this influences both the rate of reporting of potential adverse events and the
acceptability of them.

Passive reporting is particularly deficient in that only potential events in vaccine recipients are
reported, and even then probably only a minority. The number of people who receive the
vaccine in total is usually not known, nor is the rate of the event in those who did not receive the
vaccine.

Case reports and series suffer from these deficiencies as well, but in addition may be affected
by referral bias once a physician becomes known to be interested in the condition and its
alleged association with vaccine.

In analytical epidemiological studies, whether case control or cohort, there are many potential
sources of bias and confounding which make results difficult to interpret. The Institute of
Medicine, for instance, noted that studies of a potential link between sudden infant death
syndrome and diphtheria-tetanus-pertussis vaccine had initially found an increased risk but once
designs and analysis were refined the relationship was consistently protective.
The temporal relationship between receipt of the vaccine and onset of the alleged adverse event is
probably the strongest evidence for a true relationship. This can best be examined through data
linkage of routine vaccine records to routine health outcome data with the addition of demographic
information. Even in this situation careful adjustment of confounding is required, in particular
changes in the underlying rate of disease with time.
Complementary sources of information on vaccine safety come from diseases associated with
natural infection, differences between the natural agent and the vaccine, understanding of the pathophysiology of the condition concerned and background population data on rates of illness –
particularly in comparable groups to those targeted for vaccination.
The 1992 Institute of Medicine Report on Vaccine Safety found that there was inadequate evidence
to either accept or reject an association between hepatitis B vaccine and any neurological disorder
. The reasons being that:
 randomised clinical trials lack power.
 post-licensure surveillance is limited (mostly passive).
 there is competition for resource between safety assessment and control activities.
Neurologic disorders. In: Adverse events associated with childhood vaccines: evidence bearing on causality. Stratton KR,
Howe CJ, Johnston RB, eds. Institute of Medicine, National Academic Press, Washington DC, 1994.
Hepatitis B vaccine. In: Adverse events associated with childhood vaccines: evidence bearing on causality. Stratton KR,
Howe CJ, Johnston RB, eds. Institute of Medicine, National Academic Press, Washington DC, 1994: 211-35.
5
2.2.2. Vaccine safety and causality
To demonstrate a link between vaccination and the onset of an unrelated illness, certain
epidemiological criteria must be met in any epidemiological study. These criteria (known as Bradford
Hill’s criteria ) include temporality, consistency, coherence, specificity, biological plausibility and
strength of association, consistency being particularly important.
The assessment of causality is a difficult judgement. Epidemiologists have often been poor at
communicating both the method used and the conclusions from it. There is a need for simpler
messages emphasising that animal studies, case reports and series, passive surveillance and
ecological studies can only be hypothesis generating.
3.
Auto-immune diseases, including multiple sclerosis
MS is a disease of the central nervous system (CNS) characterized by the destruction of the myelin
sheath surrounding neurons. MS is a progressive, often fluctuating disease with exacerbations and
remissions over many decades usually resulting in permanent disability and sometimes death .
The cause of MS is unknown. The most widely accepted hypothesis is that MS occurs in patients
with a genetic susceptibility, and that some environmental factor or factors ‘trigger’ clinical
exacerbations. The genetic predisposition is well supported by many studies showing an increased
risk of MS in family members of cases, identical twins, certain ethnic populations and certain HLA
subtypes.
Multiple sclerosis is the commonest CNS disease in the 20-45 age group. It shows a peak incidence
around the age of 30 and this distribution has been considered to be consistent with the incubation
period distribution for an infection in adolescence such as infectious mononucleosis. The disease is
more common in females than males by a ratio of 3:1, as is seen in other autoimmune disorders.
The diagnosis is based on symptomatology consistent with plaques in at least two anatomical
locations. For this reason the diagnosis is often delayed following the first episode of demyelination.
Current therapeutic interventions involve various immune modulatory drugs.
Studies in twins show concordance in monozygous twins of around 26% but only 4% in dizygous
twins. This clearly indicates both a genetic component to the disease and, since more than 70% of
monozygous twins are not concordant, environmental factors in the aetiology. Three studies using
genomic scans to determine the location of the responsible genes were inconclusive.
There is a striking geographical pattern with disease prevalence and incidence increasing with
latitude as one moves north or south from the equator. Migrant studies demonstrate that risk of
disease changes with migration between these latitude bands and that this effect is dependent on
the age at migration.
Susser M. Falsification, verification and casual inference in epidemiology: reconsideration in the light of Sir Karl Popper's
philosophy. In: Rothman KR, editor. Casual Inference. 4th printing. Massachussets: Epidemiology Resources, Inc. 1988: 3357.
Weinshenker BG. The natural history of multiple sclerosis. Neurologic Clinics 1995; 13: 119-46.
6
Descriptive studies in Norway have clearly shown time/space clustering of residence in adolescence
as well as differing trends in disease in different geographical areas.
Environmental agents that have been proposed as aetiological agents include:
 Infectious diseases, either as triggers of auto-immunity or through persistent infection;
 diet (primarily animal fats);
 trauma to the head;
 toxic agents, such as heavy metals and organic solvents.
There is some consistency in the finding for organic solvents. However, further aetiological research
is needed to guide rational treatment as well as prevention. In particular, studies of
gene/environment interactions are required. It is proposed that well designed, multi-centre analytical
studies with sufficient power for sub-group analyses are needed.
3.1. Immunological mechanisms and molecular mimicry
Several potential mechanisms are thought to underlie autoimmune processes including multiple
sclerosis. The model of autoimmunity is one in which genetic factors – both Major Histocompatibility
Complex (MHC) and non-MHC overlap with environmental factors in contributing to the disease.
The specific mechanism in multiple sclerosis is an immunological mechanism initiated by one or
more viruses. The virus could potentially induce autoimmunity in one, or a combination of, four
ways:
 Polyclonal activation
 Infection of lymphoid cells
 Anti-idiotype
 Molecular mimicry
Polyclonal activators cause a clonal expansion of B cells leading to an anti-self response. EpsteinBarr virus or lipopolysaccharide antigen are examples of these. An alternative or complementary
pathway is through superantigen stimulation of T cells. If superantigen stimulates naive T cells then
expansion followed by apoptosis results. If the T cells are already activated, however, then
autoimmunity can result. These stimuli can act through either complement mediated cell lysis or
immune complex mediated cell injury. Polyarteritis nodosa that can be associated with hepatitis B
surface antigen is a classic example of immune complex mediated disease.
Viral infection of lymphoid cells might either be of B cells or other antigen presenting cells – leading
to up regulation of self antigens and presentation of these (for example in heat shock proteins) or T
cells leading to up regulation through cytokine production (in particular gamma interferon).
Anti-idiotype mechanisms of induction of autoimmunity might arise from a common antigen on the
surface of both virus and cell surface. Under normal circumstances, the development of anti-idiotype
antibody to the human cell antibody regulates immune response to subsequent viral challenge.
However, if the viral challenge occurs in the neonate then the protective idiotype antibody may not
be produced. The self antigen- antibody would then attach to tissues and induce an inflammatory
response.
Molecular mimicry occurs when there is sequence homology between antigen in the infecting
organism and the host. This may result from the virus inherently coding for cross-reactive
determinants or by the incorporation of host components into the virion. The classic example of this
is the antigen on the Streptococcus pyogenes that cross reacts with joint, heart and skin antigens
and can lead to rheumatic fever.
Fujinami RS. Molecular Mimicry: virus modulation of the immune response. In: McKendall RR, Stroop WG, eds.
Handbook of Neurovirology. New York: Marcel Dekker, Inc., 1994: 103-14.
7
Multiple situations are known where molecular mimicry occurs (e.g. Coxsackie B virus and
myocardial tissue, measles virus and stress proteins or myelin basic protein). If the viral antigen is
processed by antigen presenting cells and presented in association with class I MHC this results in a
cytotoxic T cell response whereas if it is class II MHC then CD4 mediated auto-immunity results.
3.2. Animal models and demyelination
The best animal model used to simulate the demyelination of multiple sclerosis is experimental
allergic encephalitis in the mouse. A neuroantigen in Freunds complete antigen is injected into the
mouse resulting in perivascular cellular infiltrates in the brain and demyelination. This results
clinically in a limp tail, hind limb paralysis and incontinence.
The responsible encephalitogenic regions of myelin basic protein differ in the rabbit, mouse and in
man. Sequence homology (molecular mimicry) has been demonstrated between the rabbit
encephalitogenic protein and the polymerase gene of hepatitis B surface antigen . It is to be noted
that hepatitis B vaccine does not contain this gene and that it is to the rabbit encephalitogenic
protein. However, injection of rabbits with the viral peptide produced myelin basic protein (MBP)
specific antibodies in a majority of rabbits, some of which developed inflammatory changes in the
brain, thus validating mimicry as a mechanism of neuro-inflammatory disease.
In the mouse, the use of a vaccinia recombinant expressing the encephalitogenic protein induced
protection against experimental allergic encephalitis (EAE). This offers a model for the primary
prevention of demyelinating diseases. Alternatively, the use of vaccinia expressing myelin proteolipid
protein (PLP) primed mice for EAE following challenge with encephalitogenic peptides derived from
PLP. Interestingly, the demyelination in these animals did not show the relapsing pattern of classical
EAE
.
To date, the search of the recombinant surface antigen gene of hepatitis B has revealed no
homology between albumin, myelin basic protein and other CNS proteins.
4.
Available data on safety of hepatitis B vaccines
4.1. Literature and clinical trials
Safety trials have been carried out in healthy individuals of all ages as well as in those at increased
risk of infection and those with impaired immunity (for example haemodialysis patients). The
recording of adverse events within trials is the responsibility of the investigator. Active surveillance is
usually carried out for the first 4 to 14 days after each dose of vaccine and unsolicited recording of
health events occurs for 30 days after each dose. Serious adverse events are recorded throughout
any study which may be up to 7 or 13 months in immunogenicity studies or longer in phase III trials.
Deaths and congenital abnormalities should be reported at any time a) if they are brought to the
attention of the investigator and b) if they are thought to be vaccine related.
A literature search found that 350 clinical trials on the safety of hepatitis B vaccines have been
conducted. From these, information is available on 230 trials involving some 18,000 subjects. In
these trials, 2,000 serious adverse events were reported, of which 111 were considered by the
investigators to be related to the HB vaccine. Interpretation of this data is difficult, however, because
Fujinami RS, Oldstone MB. Amino acid homology between the encephalitogenic site of myelin basic protein and virus:
mechanism for autoimmunity. Science 1985; 230: 1043-45.
Barnett LA, Whitton JL, Wada Y, Fujinami RS. Enhancement of autoimmune disease using recombinant vaccinia virus
encoding myelin proteolipid protein. J Neuroimmunology 1993; 44: 15-26.
Barnett LA, Whitton JL, Wang L-Y, Fujinami RS. Virus encoding an encephalitogenic peptide protects mice from
experimental allergic encephalomyelitis. J Neuroimmunology 1996; 64: 163-73.
8
of the subjectivity involved in each trial report. The sensitivity of trials for serious adverse events is
also limited beyond the period of active follow up.
4.2. Data from the USA
The Vaccine Adverse Event Reporting System (VAERS) has been fully operational in the USA since
November 1990. In 1991, one of the earliest reports to the system was from a Wisconsin physician
who reported four nurses affected by demyelination subsequent to HB vaccination (from a total
nursing population of 20,000). Examination of the VAERS data (1991) showed that rates of
neurological disorders were ten times higher in hepatitis B vaccinees than with any other vaccine.
This held true for non-demyelinating symptomatology (dizziness, paraesthesiae, and vasodilatation)
as well as clear demyelination. This increased reporting of symptoms is most likely attributable to
the fact that the majority of recipients of HB vaccine were healthcare workers who are most likely to
report possible adverse events.
An examination of demyelination events by time since vaccination did not show the expected peak in
week 1 (as observed with influenza vaccine and Guillain Barré syndrome
). More recent safety
data relating to recombinant hepatitis B vaccine in children has been published which shows no
unexpected adverse events in neonates and infants given vaccine, despite the use of at least 12
million doses in these age groups
.
A review of data from VAERS on demyelinating disease for 1991-1997 showed 68 reports of
multiple sclerosis. The mean age of the vaccinee was 35 (range 15-58 years), with 77% female and
10% male (13% did not supply this information). The median reported interval from vaccination to
onset of illness was 13 days. In addition, 49 reports of optic neuritis and 26 of transverse myelitis
were noted. When these events were analysed by dose of vaccine (1st, 2nd, 3rd) no pattern
emerged and no conclusion could be drawn from this data. In contrast, the reports of hair loss in
association with routine hepatitis B vaccination are likely to be causal – particularly in view of the
effect of re-challenge with the vaccine .
Langmuir AD, Bregman DJ, Kurland LT, et al. An epidemiologic and clinical evaluation of Guillain-Barré syndrome
reported in association with the administration of swine influenza vaccines. Am J Epidemiol 1984; 119: 841-79.
Ropper AH, Victor M. Influenza vaccination and the Guillain-Barré syndrome. New Engl J Med 1998; 339: 1845-1846.
Lasky T, Terracciano G, Magder L, et al. The Guillain-Barré syndrome and the 1992-1993 and 1993-1994 influenza
vaccines. New Engl J Med 1998; 339: 1797-802.
Niu MT, Davis DM, Ellenberg S. Recombinant hepatitis B vaccination of neonates and infants: emerging safety data from
the Vaccine Adverse Event Reporting System. Pediatr Infect Dis J 1996; 15: 771-6.
Mac Mahon BJ, Helminiak C, Wainwright RB, Bulkow L, Wainwright K. Frequency of adverse reactions to hepatitis B
vaccine in 43,618 persons. Am J Med 1992, 92: 254-6.
Shaw F, Graham D, Guess H, et al. Postmarketing surveillance for neurologic adverse events reported after hepatitis B
vaccination. Experience of the first three years. Am J Epidemiol 1988; 127: 337-52
Wise RP, Kiminyo KP, Salive ME. Hair loss after routine immunizations. J Am Med Assoc 1997; 278: 1176-8.
9
4.3. Data from Canada
In Canada, some 5 million doses have been delivered in vaccination programmes targeting preadolescents, infants and risk groups (varying by province and over time). Between 1990 and 1998,
some 2,735 adverse events were reported, 65% of these by females. The distribution of events was
similar to the age distribution of vaccination. In addition, there were reports of chronic fatigue
syndrome associated with the vaccine, although these reports came from a TV programme.
Analysis of the time relationship to vaccination showed no relationship between HB vaccination and
MS. Several specific studies of the putative relationship with multiple sclerosis are planned in
Canada.
4.4. Data from Italy
Hepatitis B vaccination is obligatory in Italy, resulting in infant vaccination coverage in excess of
90%. In total, over 15 million doses of vaccine have been delivered. Compensation is available for
proven vaccine induced harm and this facilitates reporting of adverse events. In total, 504 adverse
events have been reported, giving a rate of 3/100,000 doses; 19 of these involved the central
nervous system (for a rate of 1/million doses) although none were multiple sclerosis. The vaccine
manufacturer (SmithKline Beecham) operates an independent database. When this was searched,
5 patients with CNS events were found in Italy, one of whom had multiple sclerosis.
4.5. Data from Industry
Some 630 million doses of hepatitis B vaccine have been distributed worldwide, 69 million of these
in France. Based on the database maintained by the companies reports of multiple sclerosis (274 in
number) were at a rate of 0.19/100,000 doses globally and 0.59 per 100,000 in France. Examination
of the time relationship between vaccine receipt and disease onset showed no pattern to suggest
causality.
5.
Review of the epidemiological studies in progress: preliminary data and summary of
study protocols
5.1. France
An increase in reports of central nervous system demyelinating disease (CDD) possibly associated
with hepatitis B vaccine at the end of 1996 led to the initiation of two epidemiological exercises in
France.
First, the expected number of cases of CDD that would have occurred by chance in association with
vaccination was calculated based on the underlying rate of disease and the number of doses of
hepatitis B vaccine distributed in France. The observed number of cases was less than the
expected. However, because of a number of uncertainties in this calculation a case control study
was carried out.
Incident cases of CDD with their first episode of demyelination occurring within the last six months
were recruited from hospital neurology departments. Controls also came from the same neurology
departments. They were patients with other neurological conditions some of which, for example
migraine, were chronic. They were selected by the neurologists and matched on age, sex and date
Bentsi-Enchill A. Adverse events after hepatitis B vaccination. Can Med Assoc J 1992; 147: 1023-6.
Dobson S, Scheifele D, Bell A. Assessment of a universal, school-based hepatitis B vaccination program. J Am Med
Assoc 1995; 274: 1209-13.
10
of examination. Subjects with contraindications to vaccination were excluded. The vaccination status
of subjects was determined by telephone interview following a letter to the subject requesting that
they find their records and fixing a time for the telephone conversation. The preliminary results were
presented using a variety of case sub-groups and time intervals. None of these showed statistically
significant associations between HB vaccination and CDD.
There was some concern about the design of this study. In particular, the selection of controls was
less than ideal since pre-existing disease might have influenced their likelihood to be vaccinated.
Community controls would have been preferable. There was some reassurance on this point since
the hospital controls’ vaccine experience was similar to that of the general population.
5.2. UK
Analysis of a computerized, UK-based General Practitioner data base covering some 4 million
people was carried out. The database includes prescriptions for vaccination as well as disease
codes. Individuals aged 20-60 years with episodes of CDD between 1/1/90 and 1/6/97 were selected
from the database and matched by age, sex and practice to six controls each. Prior vaccination with
hepatitis B and other vaccines was examined.
Further refinements to the analysis are needed but preliminary results show no association with
multiple sclerosis but a marginal association with acute demyelination not diagnosed as multiple
sclerosis. Potential confounding and bias have been carefully addressed in this study and controlled
as far as was possible.
The study had the major advantage of prospective identification of exposure to vaccination and
outcome. Although cases had not yet been formally validated, the diagnosis was based on hospital
attendance and these records had been independently reviewed by three physicians blinded to
exposure status. There was some concern about the completeness of vaccination data since
vaccines given at Occupational Health departments, STD clinics and Travel clinics might not be
included in the GP records. However, any misclassification would be non-differential.
5.3. USA
An analysis of a research database from the Diversified Pharmaceutical Services, a pharmacy
benefits management company, was carried out in the USA. The research database was formed
from six primary healthcare programmes and records reimbursement claims for both hepatitis B
vaccination and medical treatment for demyelinating disorders. A cohort analysis was performed
defining a group of 27,229 who had claimed for hepatitis B vaccination and a group of 107,469 who
had not claimed for hepatitis B vaccination. These groups were matched on age and sex. Analysis
of rates showed rate ratios of less than one for all outcomes indicating a potential protective effect of
hepatitis B vaccination. This was interpreted as potentially due to a ‘healthy vaccinee’ effect.
Concern about the validity of both exposure and outcome measurement in this study was
expressed. In addition it was not clear that person years at risk by time interval had been calculated
appropriately.
5.4. Industry
A study has been proposed and funded by Pasteur Mérieux - Connaught to address the issue of
whether hepatitis B vaccination influences the rates and timing of onset of relapses in multiple
sclerosis. The design is a case-crossover design in which three month periods of time preceding a
relapse in the individual will be compared to three month periods when not preceding a relapse.
Thus the cases act as their own control. A pilot study found a prevalence of vaccination in a three
month period of 12%. Based on this, a sample size of 600 was estimated to detect a RR of 2.0.
11
It was noted that many of these patients might be receiving immuno-modulatory drugs which could
modify the effect of vaccination. This effect modification will have to be examined in the analysis and
to have the power to do this a larger sample size – perhaps fourfold larger – is needed.
5.5. Summary and review of the epidemiological studies
Multiple Sclerosis is almost certainly a multi-factorial condition (or conditions) in which some causes
act during or prior to adolescence in genetically susceptible individuals. The disease epitomises the
difficulties of designing studies aimed at infectious disease interplay between environmental and
genetic factors.
The presentation of the epidemiologic data generated a discussion in which following points were
made:
 The presentations highlighted the range of variables that were potential confounders in the study
of a possible association between hepatitis B vaccine and multiple sclerosis. In particular the
need to control for adolescent infections and socio-economic group.
 The interval from vaccination to onset of demyelination in the cases reported from France was
of the order of two months. It was unclear how this fitted with data suggesting that the aetiology
of MS acted some decade before the peak incidence.
 In the light of this, the possibility was raised that vaccination might only act as a trigger
precipitating MS in individuals who would inevitably develop the disease in time.
 No studies appeared to have addressed an association between natural hepatitis B infection
and MS. Such a study could possibly be performed using stored sera from completed case
control studies.
 The geographical pattern of multiple sclerosis seemed almost to be a mirror image of persistent
hepatitis B infection and this might deserve formal analysis.
 The putative association raised the question of the incidence of multiple sclerosis in the highly
vaccinated, and high latitude, Eskimo population in Alaska.
 The question was raised whether it is biologically plausible that no association exists between
natural hepatitis B carriage (where there are massive amounts of HBsAg in the circulation) and
MS whereas there is a putative association with the injection of relatively small quantities of
surface antigen as vaccine. The immunologists felt that this was plausible given the potential for
large dose tolerance induction. In addition, natural replicative infection is quite different from
vaccination with only part of the non-replicative antigenic repertoire of the virus.
In general, the epidemiologic studies presented at the meeting were preliminary and had several
limitations, especially statistical power.
To date:
 None of the studies have findings which are statistically significant.
 None of the studies are as yet published in peer review journals.
 None of the studies have (yet) used Guidelines for Future Analytical Research .
 No attempt has yet been made to calculate the cost benefit situation in terms of the upper 95%
confidence interval on any estimate of effect, should an association exist.
Two publications from the late 80s did not find any association between vaccine and CNS
disorders21 22.
Riise T, Wolfson Ch. The epidemiologic study of exogenous factors in the etiology of multiple sclerosis. Guidelines for
future analytic research. Neurology 1997; 49 (suppl. 2): S1-S82.
12
6. Conclusions and consensus
There is no doubt that currently available hepatitis B vaccines are highly efficacious in inducing
protective immunity; the hepatitis B vaccine has protected over 200 million people worldwide from
hepatitis B virus infection, a serious and untreatable disease that can lead to cirrhosis and liver
cancer. For the 4 million children born in the US each year, universal hepatitis B immunization will
prevent 7,500 deaths from cirrhosis and liver cancer. Recently, it has been suggested that
vaccination against hepatitis B might cause or precipitate CNS demyelinating disorders including
MS.
As with all medicines, regulatory authorities assess the safety and effectiveness of vaccines
throughout their development and continue to monitor both safety and effectiveness when used in
the population at large. When concerns arise, they are pursued with in-depth studies to determine
whether they represent any risk to the population.
A scientific group of experts met at WHO to review the epidemiology and current understanding of
the pathogenesis of MS. Current understanding of these diseases indicates a multi-factorial
pathogenesis and the contribution of both genetic and environmental factors. Numerous
environmental factors including infectious agents have been evaluated in studies conducted
throughout the world. Both the age distribution of onset of disease and migration studies suggest
that exposure to an environmental agent in early childhood or adolescence contributes to the
pathogenesis of MS with a 10-15 years interval to onset of disease. An increase in risk with increase
in latitude has been consistently demonstrated, but this is a marker for an as yet unidentified factor.
There are three hypotheses that could explain the observed cases of demyelinating disorders
following HB vaccine:
1. Coincidence, due to the large number of HB vaccine doses administered, many of them in age
groups where symptoms of MS first occur.
2. “Triggering”: An increased risk of symptomatic demyelination following HB vaccine which would
act as a “trigger” in individuals predisposed to develop MS or central nervous system (CNS)
demyelinating diseases. These individuals would have developed demyelination with or without
an altered natural history following some immunologic or other precipitating factor.
3. A true causal relationship between HB vaccination and MS or other CNS demyelinating disease.
Evidence to support the first hypothesis includes the fact that no statistically significant association
was found between hepatitis B vaccine and MS in the limited studies conducted to date. Further, the
age and sex distributions of MS cases reported through spontaneous reporting systems match the
recognised age and sex distribution of MS cases that preceded the use of the vaccine and are not
correlated with vaccine administration.
In support of the second hypothesis of an increased risk of MS following HB vaccination seen as a
precipitating factor is that some studies have shown slightly elevated odds ratios although these
were not statistically significant. Evidence inconsistent with this second hypothesis is the observation
that no increased risk was found in another study. No tangible evidence was presented for the
biologic plausibility of any association.
There is no evidence whatsoever of a causal link between hepatitis B virus infection and CNS
demyelinating disorders including multiple sclerosis. Additional epidemiological and immunological
research is ongoing or planned to further examine any association between vaccination, including
hepatitis B, and CNS demyelinating disease. Altogether, evidence in favour of an increased risk
following vaccination is weak and does not meet the criteria for causality.
The group reached the following consensus:
13

Epidemiological studies have identified that transmission of hepatitis B virus occurs through
contact with contaminated blood and other clinical materials, sexual transmission, child to child
transmission and through household contacts. Yet in almost all countries studied, approximately
1/3 of hepatitis cases occurs in people who are not included in these identified high risk groups.
Strategies that target only groups at high risk of hepatitis B virus infection have failed, therefore
strategies that protect the population at large are encouraged.

Those who have already been vaccinated should be reassured that no studies demonstrate that
they are at increased risk of developing MS.

MS does not occur in children <1 year of age. There is no reason to believe that infant
immunization with hepatitis B vaccine will change this fact.

MS is a very rare disease in children and adolescents. Amongst the cases of MS that have been
reported to surveillance systems, there is no evidence that MS was caused by the hepatitis B
vaccine in these cases.

By better understanding the pathogenesis and aetiology of MS and related demyelinating
disorders one would be in a position to identify factors that influence its initiation and outcome.
Therefore, research into the causation and immunopathogenesis of CNS demyelinating
disorders including MS should be encouraged and supported.

Similarly, the influence of HB vaccination on immune functions, including immunological
memory, should be further investigated. An important aspect of immunological research
concerns the availability of biological material including T and B lymphocytes, monocytes, serum
and plasma which has been stored properly. Therefore a close collaboration between
epidemiologists and clinical researchers is essential. This also implies close liaison with
regulatory authorities.

Any future or ongoing study of demyelinating disorder and HB vaccine should take advantage of
the large experience gathered to date on conducting MS aetiologic studies and the Guidelines
developed. The vaccine and MS communities would benefit from jointly organizing the best
possible case-control study on aetiology of MS. Such a study would examine a wide range of
hypotheses (though each being as specific as possible); including different viral, bacterial and
parasital infections, environmental factors, and vaccinations, and control for various potential
confounders.

Principal investigators of existing studies should standardise methodology to facilitate future
meta-analysis. In addition, validation of vaccination exposure should be improved by obtaining a
copy of the vaccination certificate and/or serology for hepatitis B surface antibody. Researchers
should also collect better information on potential confounders.

The creation of an European Union vaccine pharmacovigilance and vaccine safety studies
network would improve signal generation and hypothesis testing.

The data available to date, although limited, does not demonstrate a causal association between
HB immunization and CNS demyelinating diseases, including MS. No evidence presented at
this meeting indicates a need to change public health policies with respect to HB immunization.
Therefore, based on demonstrated important benefits – including the prevention of cirrhosis and
cancer, and a hypothetical risk, the group supports the WHO recommendations that all
countries should have universal infant and/or adolescent immunization programs and continue
to immunize adults at increased risk of HB infection as appropriate.
ANNEX 1
14
TECHNICAL CONSULTATION ON SAFETY OF HEPATITIS B VACCINES – PARTICIPANTS
VHPB Core members
Dr Peter Grob
University Hospital
Department of Medicine
Häldeli Weg 4
8044 Zurich
Switzerland
Tel: 41 1 634 2861
Fax: 41 1 634 2901
e-mail:peter.grob@dim.usz.ch
Dr Mark Kane
Director
Bill and Melinda Gates
Children’s Vaccine Programme
PATH
4, Nickerson Street
Seattle WA 98109
USA
Tel: 1 206 285 3500
Fax: 1 206 285 6619
e-mail:mkane@path.org
Dr Johannes Hallauer
Universitätsklinikum Charite
Gesundheitssystemforschung
Schumannstrasse 20/21
10117 Berlin
Germany
Tel: 49 30 2802 5919-2003
Fax: 49 30 2802 8017
e-mail:johannes.hallauer@charite.de
Dr André Meheus
University of Antwerp
Epidemiology & Community Medicine
Universiteitsplein 1
B-2610 Antwerp
Belgium
Tel: 32 3 820 2523
Fax: 32 3 820 2640
e-mail:engelen@uia.ua.ac.be
Dr Colette Roure
WHO Regional Office
Regional Adviser EPI
8, Scherfigsvej
2100 Copenhagen O
Denmark
Tel: 45 39 171534
Fax: 45 39 171851
e-mail:cro@who.dk
VHPB Standing Advisers
Dr Paolo Bonanni
University of Genoa
Preventive Medicine
Via Pastore 1
16132 Genoa
Italy
Tel: 39 010 353 8530
Fax: 39 010 353 8407
e-mail:bonanni@dsp.igiene.unifi.it
Dr José De La Torre
Ministerio de Sanidad y Consumo
Dept de Enfermedades Trasmisibles
Paseo del Prado N 20
Madrid
Spain
Tel: 349 1 596 4184
Fax: 349 1 596 1548
e-mail:sesteban@msc.es
or
University of Florence
Hygiene & Epidemiology Public Health Department
Viale Morgagni 48
50134 Florence
Italy
Tel: 39 055 411113
Fax: 39 055 4222541
Dr Nicole Guérin
129, Avenue Gide
94270 Le Kremlin-Bicetre
France
Tel: 33 1 465 83280
Fax: 33 1 465 83280
15
Dr Daniel Lavanchy
World Health Organization
Division of Communicable Diseases
20, via Appia
1211 Geneva
Switzerland
Tel: 41 22 791 2656
Fax: 41 22 7914878
e-mail:lavanchyd@who.ch
Dr Harold Margolis
Centers for Diseases Control
Hepatitis Branch
1600, Clifton Road, NE – Mailstop A-33
30333 Atlanta Georgia
USA
Tel: 1 404 639 2343
Fax: 1 404 639 1563
e-mail:hsm1@ciddvd1.em.cdc.gov
Prof Dr Georges Papaevangelou
6, Vas Georgiou Sqr Psychiko
15452 Athens
Greece
Tel: 30 1 671 3690
Fax: 30 1 6747 221
e-mail:gpapaeva@cc.uoa.gr
Prof Dr Daniel Shouval
Hadassah University Hospital
Director, Liver Unit
POB 12000
91120 Jerusalem
Israel
Tel: 972 2 6777337
Fax: 972 2 6420338
e-mail:shouval@cc.juji.ac.il
Rapporteur
Dr. Andy Hall
London School of Hygiene & Tropical Medicine
Keppelstreet
London WCIE 7HT
UK
Tel : 44/171/9272272
Fax :44/171/4364230
e-mail:andy.hall@lshtm.ac.uk
Invited Guests
Dr Monique Breteler
Erasmus Universiteit
Chief Neuro-Epidemiology
POB 1738
3000 DR Rotterdam
Netherlands
Tel: 31 10 408 7489
Fax: 31 10 436 5933
e-mail:breteler@epib.gff.eur.nl
Dr Vito Caserta
Division of Vaccine Injury Compensation
Chief Medical Officer – Medical Analysis Branch
294, Mary Frances Court
21703 Frederick, Maryland
USA
Tel: 1 301 443 4956
Fax: 1 301 4431933
e-mail:Vcaserta@hrsa.dhhs.gov
Dr Robert Chen
CDC
Chief, Vaccine Safety and
Development Activities
MS-61 National Immunization Program
GA 30333 Atlanta
USA
Tel: 1 404 639 8778
Fax: 1 404 639 8834
e-mail:rtc1@cdc.gov
Dr John Clemens
WHO
Expanded Programme
20, Avenue Appia
1211 Geneva 27
Switzerland
Dr Philippe Duclos
WHO
Medical Officer EPI
20, Avenue Appia
1211 Geneva 27
Switzerland
Tel: 41 22 791 4527
Fax: 41 22 791 4193
e-mail:duclosp@who.ch
Dr Robert Fuijinami
Utah University
Department of Neurology
50, North Medical Drive
UT 84132 Salt Lake City
USA
Tel: 1 801 585 3305
Fax: 1 801 585 3311
e-mail:robert.fujinami@hsc.utah.edu
Dr Bruce Gellin
Vanderbilt University Medical Center
Dr Pilar Gavino
WHO
16
Department of Preventive Medicine A-1124
37232-263 Nashville, Tennessee
USA
Tel: 1 615 343 6306
Fax: 1 615 343 8722
e-mail:bruce.gellin@mcmail.vanderbilt.edu
Viral Diseases Unit
20, Via Appia
1211 Geneva
Switzerland
Tel: 41 22 791 2844
Fax: 41 22 791 4878
e-mail:gaviniop@who.ch
Dr Olivier Gout
Hôpital de la Salpétrière
47, boulevard de l’Hôpital
Paris Cedex 13
France
Tel: 33 1 421 61762
Fax: 33 1 421 61965
e-mail:ogout@fo.rothschild.fr
Dr Neal Halsey
John Hopkins
Department of International Health
615 N Wolfe Street
MD 21205 Baltimore
USA
Tel: 1 410 955 6964
Fax: 1 410 502 6733
e-mail:nhalsey@jhsph.edu
Dr Roland Liblau
Hôpital de la Salpétrière
Laboratoire de Neuro-Immunologie
47, Boulevard de l’Hôpital
75013 Paris
France
Tel: 33 1 421 62141
Fax: 33 1 421 62149
e-mail:106063.1005@compuserve.com
Dr Irina Lytkinia
Sanitary Epidemiological Surveillance
Head of the Department
4/9 Grafski per
129626 Moscow
Russia
Tel: 7 095 287 4614
Fax: 7 095 287 0620
Dr Hugh Nicholas
Department of Health
Communicable Diseases Branch
Wellington House
133-155 Waterloo Road
SE1 8UG London
United Kingdom
Tel: 44 171 972 4479
Fax: 44 171 972 4559
Dr Trond Riise
University of Bergen
Department of Public Health
Bergen
Norway
Tel: 47 55 586 113
Fax: 47 55 586 105
e-mail:trond.riise@isf.uib.no
Dr Michael Scholtz
WHO
Executive Director Health Technology and Drugs
20, Via Appia
1211 Geneva 27
Switzerland
Tel: 41 22 791 4889
Dr Miriam Sturkenboom
Erasmus Universiteit
Epidemiology and Biostatistics
Room 2136 – POB 1738
3000 DR Rotterdam
Netherlands
Tel: 31 10 408 7482
Fax: 31 10 436 5933
e-mail:sturkenboom@epib.fgg.eur.nl
Dr Emmanuel Touzé
INSERM
Hôpital de la Salpétrière
F-75013 Paris
France
Tel: 33 1 421 62540
Fax: 33 1 421 62541
e-mail:e.touze@waradoo.fr
Dr Robert P Wise
Centre for Biologicals Evaluation & Research
Food and Drug Administration
1401 Rockville Pike
MD 20852-1448 Rockville
USA
Tel: 1 301 827 6089
Fax: 1 301 827 3529
e-mail:wise@cber.fda.gov
Dr Hanspeter Zimmerman
Federal Office Public Health
Division Epidemiology Infectious Diseases
CH-3003 Bern
Switzerland
17
Tel: 41 31 323 8710
Fax: 41 31 323 8795
e-mail:hans-peter.zimmermann@bag.admin.ch
Pharmaceutical Industry
Dr Yuri Danilevski
Manager Merck & Co Inc
Radisson Hotel, South Wing
2, Berezhkouskaya Nar
121059 Moscow
Russia
Tel: 7 095 941 8256
Fax: 7 095 941 8276
e-mail:yuri_danileskiy@merck.com
Dr Bernadette Hendrickx
SmithKline Beecham
rue de l’Institut 89
1330 Rixensart
Belgium
Tel: 32 2 656 8111
Fax: 32 2 656 8149
e-mail:hendrick@sbbio.be
Dr Luc Hessel
Pasteur Mérieux MSD
Halle Borie
8, rue Jonas Salk
69367 Lyon Cedex
France
Tel: 33 1 443 02001
Fax: 33 1 452 57367
Dr Joseph Y Lin
Merck Vaccine Division
PO Box 4, WP37B-318
West Point PA 19486-0004
USA
Tel: 1 215 652 7991
Fax: 1 215 652 8517
Dr Elisabeth Loupi
Pasteur Mérieux Connaught
58, Avenue Leclerc
BP 7046
69348 Lyon Cedex 07
France
Tel: 33 472 737080
Fax: 33 478 873037
Dr Patricia Sadier
Pasteur Mérieux Connaught
58, Avenue Leclerc
BP 7046
69348 Lyon Cedex 07
France
Tel: 33 472 737014
Fax: 33 478 873037
e-mail:psaddier@fr.pmc-vacc.com
Dr Robert Sharrar
Merck
Mailstop BLB 30
POBox 4
PA 19446 West Point
USA
Tel: 1 610 397 2868
Fax: 1 610 397 2328
e-mail:robert_sharrar@merck.com
Dr Bernard Soubeyrand
Pasteur Mérieux MSD
Halle Borie
8, rue Jonas Salk
69367 Lyon Cedex 7
France
Tel: 33 472 804281
Fax: 33 472 804409
Dr Attila Szilagyi
Merck Vaccine Division
POBox 4WP
37T 2D2 West Point, PA19464
USA
Tel: 1 215 652 2602
Fax: 1 215 993 3835
e-mail:attila_szilagyi@merck.com
Dr Dirk Teuwen
PM-MSD
Av Jules Bordet
1140 Brussel
Belgium
Tel: 32 2 726 9584
Dr John Weil
SmithKline Beecham
rue de l'Institut 89
1330 Rixensart
België
Tel: 32 2 656 9738
Fax: 32 2 656 9058
e-mail:weil@sbbio.be
18
VHPB Secretariat
Dr Pierre Van Damme
University of Antwerp
Department of Epidemiology and
Community Medicine
Universiteitsplein 1
B-2610 Antwerp
Belgium
Tel: 32 3 820 2538
Fax: 32 3 820 2640
e-mail:pvdamme@uia.ua.ac.be
Mr Gino Verwimp
University of Antwerp
Department of Epidemiology and
Community Medicine
Universiteitsplein 1
B-2610 Antwerp
Belgium
Tel: 32 3 820 2515
Fax: 32 3 820 2640
e-mail:verwimp@uia.ua.ac.be
Mrs Emmy Engelen
University of Antwerp
Department of Epidemiology and
Community Medicine
Universiteitsplein 1
B-2610 Antwerp
Belgium
Tel: 32 3 820 2523
Fax: 32 3 820 2640
e-mail:engelen@uia.ua.ac.be
19